Forum changes: Editing of posts has been turned off until further notice.

Main Menu

Design Philosophy: Success & Failure

Started by ynnen, July 25, 2008, 04:36:45 PM

Previous topic - Next topic


Working with a variety of different RPG engines over the last few weeks has lead to some interesting and insightful conversations with co-workers and gaming buddies. One that really struck a cord with me was a fragment of a conversation between a developer and an RPG enthusiast. This conversation focused on one key aspect of the game experience: Success and Failure.

The core of this discussion is the examination of:
- How often you need to Succeed at task resolution in a game for it to be fun?
- How often can Failure occur before it deflates the experience?

A long time ago, one of my best friends (someone I respect as an indie RPG/design savant) and I discussed a concept we termed Binary Outcomes. This concept asserts that for any task, you need to essentially feel the outcome could reasonably be a Success or a Failure... From swinging a sword to charming a princess to jumping across a chasm -- both Success and Failure need to feel possible. This doesn't necessarily mean that these things are all 50/50 coin flips, but even if it's tilted 75/25 in either direction, ultimately you won't necessarily be surprised if either Success or Failure occurs.

The further to each extreme, the less likely one type of outcome becomes. If Success is 90% certain, does the outcome really need to be tested against a resolution mechanic? Sure, there's a 10% chance of failure -- but come on, 90% is darn proficient. Failure no longer feels like a viable outcome. In this scenario, when failure occurs, it's deflating and even more disappointing than in a scenario where the chance of failure may have been much, much greater.

So this leads to some interesting questions regarding game design and setting the expectations for the game experience:

1) How often should Success occur? For a generic situation where the task is of Average difficulty, being performed by someone of Average ability - what should their success rate be based on?

2) Are there practical limits to Success or Failure? At either extreme, are there points where tasks are so simple or easily within one's abilities that no roll should be required? Conversely, a task so beyond their abilities that no possibility for success exists?

3) How well does Success "scale" with skill and experience? How often should a poorly trained, feeble swordsman successfully land a strike on his equally incompetent foe? Conversely, how often should an expertly trained, physically imposing swordsman successfully land a strike on his equally proficient foe? Are these comparable situations?

4) How does "specialisation" affect perceptions of Success? I think one of the goals in an RPG system is to allow characters to feel they have a niche, a specialty that sets them apart from others. If you're the "archer" you should feel that you have a significant advantage over the others when it comes to archery. If you're the mage, there's an expectation that you should have a pretty good chance to showcase your talent by successfully casting spells. But to what extreme? How much more often should success occur (or conversely, how seldom should complete failure occur) when you're performing your specialty?
if life is a game, i need new dice.


The answer to all these questions depends on how you approach the game. If you're trying to realistically simulate the odds using dice as a probability measure the answers are going to be different than if you're using the dice as a way to create and resolve events of dramatic tension or as an abstracted or unrelated element of purely tactical consideration.

Have you read GNS and Other Matters of Role-playing Theory or examined the Big Model (via Simulationism: The Right to Dream, Gamism: Step On Up, Narrativism: Story Now) in the Articles section? Much of what is discussed in there might help answer your questions, depending on what sort of game you are looking at playing or designing.

Which brings us to the subject-at-hand: what do you want the game you are asking this question in reference of to do in play? What do you want to emphasize and what do you want to avoid? How do you envision a session of play, what events take place during such?

Because the thing is, the answers to those questions can easily change depending on the game, and there are no "wrong" or "invalid" answers if the answers make the game do what it was designed to do.
Rev. Ravenscrye Grey Daegmorgan
Wild Hunt Studio


I have indeed ready the GNS theory articles and many of the other design/theory articles and threads. This is a loaded question since, as you point out, there is no right or wrong answer. Or, you can only get close to right or wrong answers once you've placed this within the context of a setting and game premise.

These are "meta" level design questions I'm simply seeking input and feedback on, leveraging the varied experience and POVs that the Forge forumites have to offer. I'm also hoping there may be a kernel of inspiration in some of the responses that lead me to ponder other meta-level design questions, or help me tighten my focus and drill down to the micro-stage at some point.
if life is a game, i need new dice.

Marshall Burns

I'm working on an article about this kind of thing, but it's not complete.  However, here's a few of my opinions (which will be elaborated on in the article, and here too, if you'd like).

Don't ask, "How often should success occur?"

Instead, ask these questions:
1.  "What is supposed to be fun about playing this game?"  This is the most important thing.  You must pin down precisely where you want the fun to come from.  GNS are the gateway to this, but they are not the answer you need.  You need someting specific; GNS are (deliberately) general.

2. "What are the literary conceits, conventions, and motifs in the kind of narrative experience I want the game to produce?"  Note that's a small "n" on narrative; I'm not talking about Narrativism, I'm talking about a narrative, a series of events, which is something that all roleplaying games have.

3. "When playing the game, who has authority over what?  When might that authority change hands?  Where might authorities intersect?"  Note that System itself can be given authority.  Where authorities intersect, that's where you need resolution.

4. "When authorities intersect, by what criteria do I weigh them to see which prevails?"  There's a LOT of criteria you can judge on here.  Character competence is only ONE of these.  Motivations, emotions, who's got more at stake, descriptive qualities, literary convetions, style, voluntary acquiescence... The list can go on, and on.  There's more waiting to be invented, even.  And you don't have to use just one (I can't think of any game that does!); mix and match them, based

5.  "In what terms do I want to express the weighing of authority?"  This is where you start thinking about DFK, and stats, and all those fiddly bits.

6.  "In what terms do I want to express the results of resolution?"  This is a crucial step that a lot of people miss. There's a thing in RPGs that I'm calling "Expression" in my article-thing, and by it I mean the terms we use to describe in-game stuff in operational terms.  To put it another way, Expression is the System of Color.  Expression is not always numerical, or even objective!  Qualitative, subjective Expressions are not only possible, but integral to roleplaying.  I mean, it can very easily become important whether the gun is in your right or left hand, or which side of the street you parked your car on, etc.  As has been said before, in an RPG, the fiction is part of the rules.

I hope this helps.  Feel free to ask me to elaborate on any of it!


Your questions are deeply simulationist. I spent a lot of time pondering similar questions and working up a system to handle difficulty, success and failure in a wide variety of circumstances. Ultimately, I gave up on it. There are simply far too many variables and complex interactions to satisfactorily create a model (simulation) in any believable fashion. More importantly however, I realized that simulation wasn't really in the sweet spot I'm going for.

I make a careful distinction between character success and player success, and I approach this kind of thing from the perspective of player success. As a player, I am "succeeding" if I'm having fun at the table. Now, my definition of "fun" is tied to my style of roleplay and is probably different than yours. For some people, character success doesn't equate to fun; it may be more enjoyable/satisfying if their character fails in a dramatic, spectacular fashion that resolves a major issue for their character. Your questions imply that fun for you is simulationism, and that's fine -- but if you're going to tackle a very difficult problem be sure it's a worthwhile problem for you to solve in the context of fun and satisfying play.


Adam Dray

I've personally found a certain kind of play that is fun whether my character succeeds or fails. The trick is making failure as interesting as success. The quick way to achieve this using just about any ruleset is to avoid rolling the dice when the failure outcome isn't interesting. Additionally, make each roll count for something personal. That is, maybe you're not rolling to see if you succeed but instead you're rolling to see if you can avoid some personally unpleasant side-effects of winning.

Another trick I use is a GM bribe. Here's what I did using D&D 4e. We played "The Scryer," which is my D&D 4e interpretation of HBO's "The Wire." Every player created a "temptation" for his character. Each temptation was some bad thing that the character wanted, but was trying to avoid: alcohol, greed, vanity, glory over team, and so on. During play, whenever a character would fail some important roll, I'd offer them a bonus that would put them over the edge, but the bonus came with a condition. For example, Harva's temptation is hedonism. She is trying to prevent her informant from getting caught by some bad dudes. She fails her roll by, say, 4 points. I say, "I'll offer you a +4 to that roll if, after you save your informant, you get depressed that you let the informant get caught in the first place and you end up at the bar getting really drunk instead of going back to help the team." Note that it isn't just getting drunk, but rather getting drunk with a potentially stick consequence. Players eat this up! Even when a player passes up the bonus, it's because the cost was too high and they've made a cool statement with their failure about who they are.

Adam Dray /
Verge -- cyberpunk role-playing on the brink
FoundryMUSH - indie chat and play at 7777


It is really a question of how you interpret failure. The best systems are the ones that try to create success in the story or experience even through failure of an action. I think a really good example of this is Shock: (I'm a merciless pusher for the damn thing it seems). You don't just roll against a score, you roll against a person (your Antagonist). Thus, any situation where there is no conflict between you and the guy keeping you down is automatically successful as long as it is reasonable to achieve. Even a mundane thing, if the Antag wants to oppose it and comes up with a good reason why, can be rolled for. The real trick isn't just that though, it's that you aren't just rolling against each other, because that would create a zero-sum game, which always results in at least one guy getting the short end of the stick. What you do instead is each roll for non-mutually exclusive goals, and then get extra dice to roll against their action, so thus you can theoretically both win or loose. This means that all conflicts are inherently dramatic, because only dramatic situations where there is direct conflict result in rolls, and there is the potential even then for mutual success (the you interrogate the guy and get information from him, but you kill him in the process and thus have to flee from murder charges).

I always thought that was a fairly effective system.

I think political and social philosophy and science are helpful here, actually (they involve the study of the interaction of people in groups, and preferably obtaining the best outcome in the grand scale of things). Utilitarianism is a philosophy that concludes that the purpose of the Institution (the State, or in our case the Game System) is to arrive at the total greatest happiness/enjoyment/satisfaction/fulfillment for all members of the group. (E.G. You go see a movie with your friends but can't decide what to watch. You've already seen movie Y and haven't seen movie X, so you want to go see movie X. Everyone else hasn't seen either movie, and would much prefer movie Y. The total enjoyment resulting from seeing movie Y will thus be greater than movie X, because you are one and they are many, and your probably not going to be loosing a leg for seeing movie Y so it doesn't cause you much pain.) I think this is very apropos for these things, because what we're talking about is bringing enjoyment to the people playing. Nash Equilibrium theory is also useful I think, although both these things cover much wider issues of play than just this one little thing we're talking about (though the nature of success/failure is one of the most fundamental pieces of rpg gameplay; it is, after all, what differentiates it from simple improvisational acting).


I think you have a good point, about failing with a 10% chance of failure being worse than a 50% chance of failure. I think this depends on whether everything hinges on a single roll or whether success is determined by multiple rolls. In combat where you're swinging multiple times and have a 5% chance of failure, then even missing 1 or 2 times isn't a big deal since you know you'll succeed at your longer term goal. Having everything hinge on one roll which you know you should win at (90% success) and you fail it really sucks. Goes the other way though if you only have a 10% chance of success and you actually do succeed.